Donezo InboxSign in

Home / Research

The Price of AI Functionality in Email Productivity Software

Updated October 31, 2026. Working paper, version 1.0, October 31, 2026. A preprint that has not been peer reviewed.

A study of how 12 email software vendors price their subscriptions and which plans include AI features, by Ankur Shrestha. It reads each vendor's own pricing page, shows the data behind every number and states its limits. It is a working paper (version 1.0) that has not been peer reviewed. The author runs Donezo Inbox, which sells one of the products in this category; that product is left out of the sample.

What does it ask?

How do subscription pricing structures and AI feature access differ across selected email productivity software products? The study describes what vendors publish. It does not explain why they price as they do, and it does not claim to represent the whole market. It is exploratory: the plan for the analysis was written after the data were collected, so it was not pre-registered.

Who wrote it?

Ankur Shrestha, Donezo Inbox, United States. ORCID: 0009-0009-6036-5849. There are no other authors.

How was it done?

  • The study looks at plans: a named subscription tier of one vendor's product. Prices, billing terms, seat rules and AI feature statements come from each vendor's own pricing page, with a source link and a date for every plan.
  • The sample is hand-picked from the records of a commercial price index, so it is not a random or representative sample of the market. The author's company's own product is deliberately excluded to avoid a conflict of interest.
  • Each plan was checked again against the vendor's page on October 31, 2026. A plan whose price could not be confirmed is flagged and left out of the main results.
  • Prices are list prices in US dollars. Monthly, annual and two-year billing are never mixed: annual prices are shown per month only among annual plans, and a monthly-versus-annual comparison is made only where a vendor shows both.
  • A plan counts as including AI features only when the vendor's own page says so. If a page does not say, the paper records "not stated", not "no AI".

What data does it use?

30 plan listings from 12 vendors; the main results use the 24 plans whose prices were confirmed, from 10 vendors; 6 plans were left out because their prices could not be confirmed. The plan-by-plan data, data dictionary, analysis scripts and outputs are in the dataset, doi:10.5281/zenodo.23841631, under a Creative Commons Attribution 4.0 licence. The saved vendor pages themselves are withheld while each vendor's terms of use are checked; source links, capture dates, checksums and short quotes are included.

What did it find?

What the paper found (version 1.0; hand-picked sample, US-dollar list prices)
FindingResultPlans (vendors)
Price per seat per month on annual billing$4.17 to $64.00; median $25.0014
Annual billing compared with monthly billing, where a vendor shows both20.0% cheaper at the median11 (5 vendors)
Plans that say they include AI features16 of 2424 (10 vendors)
Plans that do not say either way4 of 2424 (10 vendors)

The paper found no pair of adjacent tiers where the price step can fairly be read as the price of AI access, so it reports no AI price premium. That is a finding about whether plans can be compared, not evidence that AI is free or already priced in.

What are its limits?

  • The sample is hand-picked and small. It says nothing about the market as a whole.
  • Prices are vendors' list prices on one date. They are not transacted prices, and promotions, regional prices and negotiated prices are not covered.
  • Whether a plan includes AI features is the author's judgement under written rules; a second person has not yet checked them.
  • The vendors were not consulted and have not reviewed the paper.
  • The author's company sells a product in this category. Read the paper with that in mind.
  • The paper has not been peer reviewed.

Where can you get it?

How do you cite it?

Ankur Shrestha (2026). The Price of AI Functionality in Email Productivity Software: A Descriptive Analysis of Subscription Tiers and Feature Bundling (Working paper, version 1.0). Zenodo. https://doi.org/10.5281/zenodo.23841609

BibTeX

@misc{shrestha2026aipricing,
  author    = {Shrestha, Ankur},
  title     = {The Price of AI Functionality in Email Productivity Software: A Descriptive Analysis of Subscription Tiers and Feature Bundling},
  year      = {2026},
  month     = oct,
  note      = {Working paper, version 1.0. Preprint, not peer reviewed},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.23841609},
  url       = {https://doi.org/10.5281/zenodo.23841609}
}

Competing interests, funding and AI assistance

Competing interests. The author is affiliated with Donezo Inbox, which develops and sells an email client that is one of the products in this category. The company also publishes the Email Software Value Index, a commercial price comparison from whose records the sample was drawn. The company's own product was deliberately excluded from the sample, the paper's dataset is separate from the commercial index, and the vendors were not consulted. The two have different purposes: the index helps people compare prices today, and the paper is a study of how those prices are structured.

Funding. No external funding.

Use of AI tools. An AI system, Claude (Anthropic), was used under the author's direction for writing and debugging the analysis scripts, mapping vendor statements to the coding rules, searching for and summarising the literature, drafting and editing the text, and building the consistency checks. It is not an author. The author reviewed the output and is responsible for the content.

Version history

  • Version 1.0, October 31, 2026: first public version. Each later version will get its own PDF and DOI and be listed here.